Non-Hodgkin's lymphoma (NHL) is the fifth most common type of cancer in the United States, and diffuse large B cell lymphoma (DLBCL) is its most common subtype. Until recently DLCBL was rapidly and inevitably fatal, but the new standard of treatment with multi-drug immuno-chemotherapy combinations makes this form of cancer potentially curable in 40% of patients. The last hope for the majority for whom standard treatment will fail lies in more radical therapies such as bone marrow transplant. These alternate treatments are more risky, especially for patients already weakened by a failed attempt at standard immuno-chemotherapy. There is thus a great need for information that could allow oncologists to individualize treatment and immediately begin radical therapy in those for whom standard treatment is likely to fail. We believe our non-invasive measurement of NHL tumor metabolism with magnetic resonance spectroscopy (MRS) can provide this information. During previous periods of grant support, we have obtained promising initial results showing that the pre-treatment metabolic ratio of phosphoethanolamine (Etn-P) and phosphocholine (Cho-P), normalized by the tumor content of nucleotide triphosphates (NTP), can sensitively and specifically predict treatment failure in DLBCL. The first aim for our multi-center study is to test our hypothesis that the [Etn-P + Cho-P]/NTP ratio is significantly correlated with treatment outcome in DLBCL. In addition to confirming our previous results with 31P MRS at the 1.5T field strength in an independent sample of patients, we also seek to extend these findings by studying the metabolism of DLBCL tumors with 31P at 3.0T, and also investigating absolute choline levels in DLBCL with 31P and 1H at both 1.5T and 3.0T. Our second aim is to follow up on initial indications that the significance of the correlation between the [Etn-P + Cho-P]/NTP ratio and treatment outcome is widely applicable to all forms of NHL and not merely the DLBCL subtype. We will approach this aim by accruing sufficient patients to analyze each subtype independently, using the 1.5T methods that are most directly applicable to the contemporary clinical environment. The proposed research is the final step leading a directed-therapy clinical trial and the culmination of an unprecedented program of translational research in MRS.